heterogeneous reaction of clono with tio and sio aerosol...

1

Heterogeneous reaction of ClONO2 with TiO2 and SiO2 aerosol particles: 1

implications for stratospheric particle injection for climate engineering 2

M. J. Tang1,2,7, J. Keeble1, P. J. Telford1,3, F. D. Pope4, P. Braesicke5, P. T. Griffiths1,3, N. L. 3

Abraham1,3, J. McGregor6, I. M. Watson2, R. A. Cox1, J. A. Pyle1,3, M. Kalberer1,* 4

5

1 Department of Chemistry, University of Cambridge, Cambridge CB2 1EW, UK 6

2 School of Earth Sciences, University of Bristol, Bristol BS8 1RJ, UK 7

3 National Centre for Atmospheric Science, NCAS, UK 8

4 School of Geography, Earth and Environmental Sciences, University of Birmingham, 9

Birmingham B15 2TT, UK 10

5 IMK-ASF, Karlsruhe Institute of Technology, Karlsruhe, Germany 11

6 Department of Chemical and Biological Engineering, University of Sheffield, Sheffield S1 12

3JD, UK 13

7 State Key Laboratory of Organic Geochemistry, Guangzhou Institute of Geochemistry, 14

Chinese Academy of Sciences, Guangzhou 510640, China 15

16

Correspondence: M. Kalberer ([email protected]) 17

Atmos. Chem. Phys. Discuss., doi:10.5194/acp-2016-756, 2016Manuscript under review for journal Atmos. Chem. Phys.Published: 23 August 2016c© Author(s) 2016. CC-BY 3.0 License.

2

Abstract 18

Deliberate injection of aerosol particles into the stratosphere is a potential climate 19

engineering scheme. Introduction of particles into the stratosphere would scatter solar radiation 20

back to space, thereby reducing the temperature at the Earth’s surface and hence the impacts 21

of global warming. Minerals such as TiO2 or SiO2 are among the potentially suitable aerosol 22

materials for stratospheric particle injection due to their greater light scattering ability 23

compared to stratospheric sulfuric acid particles. However, the heterogeneous reactivity of 24

mineral particles towards trace gases important for stratospheric chemistry largely remains 25

unknown, precluding reliable assessment of their impacts on stratospheric ozone which is of 26

key environmental significance. In this work we have investigated for the first time the 27

heterogeneous hydrolysis of ClONO2 on TiO2 and SiO2 aerosol particles at room temperature 28

and at different relative humidities (RH), using an aerosol flow tube. The uptake coefficient, 29

γ(ClONO2), on TiO2 was ~1.2×10-3 at 7% and remaining unchanged at 33% RH, and increased 30

for SiO2 from ~2×10-4 at 7% RH to ~5×10-4 at 35% RH, reaching a value of ~6×10-4 at 59% 31

RH. We have also examined the impacts of a hypothetical TiO2 injection on stratospheric 32

chemistry using the UKCA chemistry-climate model in which heterogeneous hydrolysis of 33

N2O5 and ClONO2 on TiO2 particles is considered. A TiO2 injection scenario with a solar 34

radiation scattering effect very similar to the eruption of Mt. Pinatubo was constructed. It is 35

found that compared to the eruption of Mt. Pinatubo, TiO2 injection causes less ClOx activation 36

and less ozone destruction in the lowermost stratosphere, while reduced depletion of N2O5 and 37

NOx in the middle stratosphere results in decreased ozone levels. Overall, no significant 38

difference in the vertically integrated ozone abundancies is found between TiO2 injection and 39

the eruption of Mt. Pinatubo. Future work required to further assess the impacts of TiO2 40

injection on stratospheric chemistry is also discussed. 41


3

1 Introduction 42

Climate engineering (also known as geoengineering), the deliberate and large-scale 43

intervention in the Earth’s climatic system to reduce global warming (Shepherd, 2009), has 44

been actively discussed by research communities and is also beginning to surface in the public 45

consciousness. The injection of aerosol particles (or their precursors) into the stratosphere to 46

scatter solar radiation back into space is one of the solar-radiation management (SRM) schemes 47

proposed for climate engineering (Crutzen, 2006). Sulfuric acid particles, due to their natural 48

presence in the stratosphere (SPARC, 2006), have been the main focus of stratospheric particle 49

injection research (Crutzen, 2006; Ferraro et al., 2011; Kravitz et al., 2013; Tilmes et al., 2015; 50

Jones et al., 2016). Very recently, minerals with refractive indices higher than sulphuric acid, 51

e.g., TiO2 and SiO2, have been proposed as possible alternative particles to be injected into the 52

stratosphere for climate engineering (Pope et al., 2012). For example, the refractive index at 53

550 nm is 2.5 for TiO2 and 1.5 for stratospheric sulfuric acid particles. If the size of TiO2 54

particles used for SRM can be optimized, it is estimated that compared to sulfuric acid particles, 55

the use of TiO2 requires a factor of ~3 less in mass (and a factor of ~7 less in volume) in order 56

to achieve the same solar radiation scattering effect (Pope et al., 2012). 57

Injecting particles into the stratosphere would increase the amount of aerosol particles in 58

the stratosphere, thus increasing the surface area available for heterogeneous reactions (e.g., 59

R1a, R1b, and R1c), whose effects on stratospheric chemistry and in particular on stratospheric 60

ozone depletion have been well documented for sulfuric acid particles (Molina et al., 1996; 61

Solomon, 1999). The background burden of sulfuric acid particles in the stratosphere, i.e. 62

during periods with low volcanic activities, is 0.65±0.2 Tg (SPARC, 2006). The eruption of 63

Mt. Pinatubo in 1991 delivered an additional ~30 Tg sulfuric acid particles into the stratosphere 64

(Guo et al., 2004) and subsequently produced record low levels of stratospheric ozone 65

(McCormick et al., 1995), in addition to causing substantial surface cooling (Dutton and 66


4

Christy, 1992). Observation and modelling studies have further suggested that, after the 67

eruption of Mt. Pinatubo, significant change in the partitioning of nitrogen and chlorine species 68

in the stratosphere occurred (Fahey et al., 1993; Wilson et al., 1993; Solomon, 1999), caused 69

by heterogeneous reactions of N2O5 and ClONO2 (R1a-R1c): 70

N2O5 + H2O + surface → HNO3 + HNO3 (R1a) 71

ClONO2 + H2O + surface → HNO3 + HOCl (R1b) 72

ClONO2 + HCl + surface → HNO3 + Cl2 (R1c) 73

Therefore, before any types of material can be considered for stratospheric particle injection, 74

their impact on stratospheric chemistry and ozone in particular has to be well understood 75

(Tilmes et al., 2008; Pope et al., 2012). 76

Heterogeneous reactions on sulfuric acid and polar stratospheric clouds (PSCs) have been 77

extensively studied and well characterized (Crowley et al., 2010; Ammann et al., 2013; 78

Burkholder et al., 2015). However, the reactivity of minerals (e.g., TiO2 and SiO2) towards 79

reactive trace gases in the stratosphere has received much less attention. For example, the 80

heterogeneous reactions of ClONO2 with silica (SiO2) and alumina (Al2O3) in the presence of 81

HCl (R1c) have only been explored by one previous study (Molina et al., 1997) in which 82

minerals coated on the inner wall of a flow tube were used. Further discussion of this work is 83

provided in Section 4.4. The lack of high quality kinetic data for important reactions impedes 84

reliable assessment of assessing the impact of injecting mineral particles into the stratosphere 85

on stratospheric ozone (Pope et al., 2012). TiO2 is an active photo-catalyst and its atmospheric 86

heterogeneous photochemistry also therefore deserves further investigation (Shang et al., 2010; 87

Chen et al., 2012; Romanias et al., 2012; Kebede et al., 2013; George et al., 2015). 88

To address these issues, in our previous work we have investigated the heterogeneous 89

reactions of N2O5 with TiO2 (Tang et al., 2014d) and SiO2 (Tang et al., 2014a) particles (R1a). 90

That work is extended here to the investigation of the heterogeneous hydrolysis of ClONO2 on 91


5

TiO2 and SiO2 (R1b) using an aerosol flow tube. There are only a few previous studies in which 92

the reactions of ClONO2 with airborne particles or droplets have been examined. For example, 93

the interaction of ClONO2 with sulfuric acid aerosol particles was investigated using aerosol 94

flow tubes (Hanson and Lovejoy, 1995; Ball et al., 1998; Hanson, 1998). Droplet train 95

techniques were used to study the heterogeneous reactions of ClONO2 with aqueous droplets 96

containing sulfuric acid (Robinson et al., 1997) or halide (Deiber et al., 2004). The interaction 97

of ClONO2 with airborne water ice particles was also examined (Lee et al., 1999). Our 98

experimental work, carried out at room temperature and at different RH, is the first study which 99

has investigated the heterogeneous interaction of ClONO2 with airborne mineral particles. In 100

the lower stratosphere into which particles would be injected, typical temperature and RH 101

ranges are 200-220 K and <40%, respectively (Dee et al., 2011). We note that while our 102

experimental work covers the RH range relevant for the lower stratosphere, it has only been 103

performed at room temperature instead of 200-220 K due to experimental challenges. 104

ClONO2 may also play a role in tropospheric chemistry (Finlayson-Pitts et al., 1989) 105

though its presence in the troposphere has not yet been confirmed by field measurements. The 106

importance of Cl atoms in tropospheric oxidation capacity has received increasing attention in 107

recent years (Simpson et al., 2015), and precursors of Cl atoms, e.g., ClNO2 (Osthoff et al., 108

2008; Thornton et al., 2010; Phillips et al., 2012; Bannan et al., 2015; Wang et al., 2016), Cl2 109

(Spicer et al., 1998; Riedel et al., 2012; Liao et al., 2014), and HOCl (Lawler et al., 2011), have 110

been detected in the troposphere at various locations. Cl atoms react with O3 to form ClO 111

radicals, which react with NO2 to produce ClONO2. The uptake of ClONO2 by aerosol particles 112

(R1b, R1c) may recycle ClONO2 to more photolabile species (HOCl or Cl2) and thus amplify 113

the impact of Cl atoms on tropospheric oxidation capacity (Finlayson-Pitts et al., 1989; Deiber 114

et al., 2004). Considering the widespread occurrence of reactive chlorine species (Simpson et 115

al., 2015) and mineral dust particles (Textor et al., 2006; Ginoux et al., 2012; Tang et al., 2016) 116


6

in the troposphere, our laboratory measurements can also have strong implications for 117

tropospheric chemistry. 118

Using the UKCA (United Kingdom Chemistry and Aerosol model) chemistry-climate 119

model, a preliminary assessment of the effect of injecting TiO2 into the stratosphere on 120

stratospheric chemistry and ozone was discussed in our previous work (Tang et al., 2014d). 121

This model was also used to investigate stratospheric ozone change due to volcanic sulfuric 122

acid particles after the eruption of Mt. Pinatubo in 1991 (Telford et al., 2009). In the previous 123

work (Tang et al., 2014d), we used the UKCA model to construct a case study in which TiO2 124

aerosols were distributed in the stratosphere in a similar way to the volcanic sulfuric acid 125

particles after the eruption of Mt. Pinatubo so that the solar radiation scattering effect was 126

similar for the two scenarios; however, the only heterogeneous reaction on TiO2 particles 127

considered was the uptake of N2O5 (R1a). Injection of solid aerosols into the stratosphere can 128

have a significant impact on ozone mixing ratios when heterogeneous reactions involving 129

chlorine are considered (Weisenstein et al., 2015). Several previous studies (Jackman et al., 130

1998; Danilin et al., 2001; Weisenstein et al., 2015) have considered the effects of solid alumina 131

particles on stratospheric chemistry; however, there is only very limited assessment of other 132

potential solid aerosol compositions (e.g., TiO2 and diamond) (Tang et al., 2014d). Here we 133

expand upon the previous literature by considering in our model a number of heterogeneous 134

reactions with new kinetic data on TiO2. In our current work the heterogeneous hydrolysis of 135

ClONO2 on TiO2 particles (R1b) has been included, using our new experimental data. The 136

changes in stratospheric ozone and reactive nitrogen and chlorine species are assessed by 137

comparing to the impact of the Mt. Pinatubo eruption. 138

2 Experimental section 139

The heterogeneous reaction of ClONO2 with aerosol particles was investigated at 140

different RH using an atmospheric pressure aerosol flow tube (AFT). In addition, its uptake 141


7

onto Pyrex glass was also studied, using a coated wall flow tube. N2 was used as carrier gas, 142

and all the experiments were carried out at 296±2 K. 143

2.1 Aerosol flow tube 144

2.1.1 Flow tube 145

A detailed description of the AFT was given in our previous work (Tang et al., 2014a; 146

Tang et al., 2014d), and only the key features are described here. The flow tube, as shown in 147

Figure 1, is a horizontally-mounted Pyrex glass tube (i.d.: 3.0 cm; length: 100 cm). The total 148

flow in the AFT was 1500 mL/min, leading to a linear flow velocity of 3.54 cm s-1 and a 149

maximum residence time of ~30 s. The Reynolds number is calculated to be 69, suggesting a 150

laminar flow condition in the flow tube. Under our experimental conditions, the entrance length 151

needed to develop the laminar flow is ~12 cm. The mixing length is calculated to be ~14 cm, 152

using a diffusion coefficient of 0.12 cm2 s-1 for ClONO2 in N2 at 296 K (Tang et al., 2014b). 153

Only the middle part of the flow tube (30-80 cm) was used to measure the uptake kinetics. 154

A commercial atomizer (Model 3076, TSI, USA) was used to generate an ensemble of 155

mineral aerosols. N2 at ~3 bar was applied to the atomizer to disperse the mineral/water mixture 156

(with a TiO2 or SiO2 mass fraction of ~0.5%), resulting in an aerosol flow of 3000 mL/min. 157

The aerosol flow was delivered through two diffusion dryers, and the resulting RH was adjusted 158

by varying the amount of silica gel in the diffusion dryers. 1200 mL/min flow was pumped 159

away through F1, and the remaining flow (1800 mL/min) was then delivered through a cyclone 160

(TSI, USA) to remove super-micrometre particles. This cyclone has a cut-off size of 800 nm 161

at a flow rate of 1000 mL/min. The aerosol flow could be delivered through a filter to remove 162

all the particles (to measure the wall loss rate), or alternatively the filter could be bypassed to 163

introduce aerosol particles into the AFT (to measure the total loss rate). Beyond that point, 300 164

mL/min was sampled by a scanning mobility particle sizer (SMPS), and the remaining 1500 165

mL/min flow was delivered into the AFT via the side arm. Mineral aerosols were characterized 166


8

online using a SMPS, consisting of a differential mobility analyser (DMA, TSI 3081) and a 167

condensation particle counter (CPC, TSI 3775) which was operated with a sampling flow rate 168

of 300 mL/min. The sheath flow of the DMA was set to 2000 mL/min, giving a detectable 169

mobility size range of 19-882 nm. The time resolution of the SMPS measurement was 150 s. 170

The bottom 30 cm of the AFT was coaxially inserted into another Pyrex tube (inner 171

diameter: 4.3 cm; length: 60 cm). A sheath flow (F2, 1500 mL/min) was delivered through the 172

annular space between the two coaxial tubes. The sheath flow has the same linear velocity as 173

the aerosol flow to minimize the turbulence at the end of the aerosol flow tube where the two 174

flows joined. Gases could exchange between the sheath flow and the aerosol flow because of 175

their large diffusion coefficients (~0.1 cm2 s-1) (Tang et al., 2014b), while aerosol particles 176

remained in the centre due to their much smaller diffusion coefficients, i.e. 10-7-10-6 cm2 s-1 177

(Hinds, 1996). At the end of the large Pyrex tube, a flow of 500 mL/min was sampled through 178

a 1/4’’ FEP tube which intruded 1-2 mm into the flow close to the wall of the Pyrex tube. This 179

gas-particle separation method enabled particle-free gas to be sampled, despite very high 180

aerosol concentrations used in the AFT. Sampling particle-free gas prevents particles from 181

deposition onto the inner wall of the sampling tube, and therefore minimizes the undesired loss 182

of the reactive trace gases (e.g., ClONO2 in this study) during their transport to the detector. 183

More detailed discussion of this gas-particle separation method used in the aerosol flow 184

experiments are provided elsewhere (Rouviere et al., 2010; Tang et al., 2012). 185

2.1.2 ClONO2 synthesis 186

ClONO2 was synthesized in the lab by reacting Cl2O with N2O5 (Davidson et al., 1987; 187

Fernandez et al., 2005). N2O5 crystals were synthesized by trapping the product formed from 188

mixing NO with O3 in large excess (Fahey et al., 1985). The synthesis and purification is 189

detailed in our previous study (Tang et al., 2014d). Cl2O was synthesized by reacting HgO with 190

Cl2 (Renard and Bolker, 1976; Molina et al., 1977). Cl2 from a lecture bottle was first trapped 191


9

as yellow-green liquid (a few mL) in a glass vial at -76 oC using an ethanol-dry ice bath. It was 192

then warmed up to room temperature so that all the Cl2 was evaporated and transferred to the 193

second glass vessel which contained HgO powders in excess and was kept at -76 oC. The glass 194

vessel containing liquid Cl2 and HgO powders was sealed and kept at -76 oC overnight. It was 195

then warmed up to room temperature to evaporate and transfer the formed Cl2O and any 196

remaining Cl2 to the third glass vial kept at -76 oC. Liquid Cl2O appeared dark reddish-brown 197

in colour. 198

The third vessel containing Cl2O was warmed up to room temperature to evaporate and 199

transfer Cl2O to the fourth vial which contained synthesized N2O5 and was kept at -76 oC. The 200

vial containing Cl2O and N2O5 was sealed and kept at -50 oC for 2-3 days in a cryostat. In this 201

work Cl2O was in slight excess compared to N2O5, and thus all the white powder (solid N2O5) 202

was consumed. ClONO2 is liquid at -50 oC, with a colour similar to liquid Cl2. The major 203

impurity of our synthesized ClONO2 was Cl2O, and the boiling temperature at 760 Torr is 2 oC 204

for Cl2O and ~22 oC for ClONO2 (Stull, 1947; Renard and Bolker, 1976). To purify our 205

synthesized ClONO2, the vial containing ClONO2 was warmed up to 5 oC and connected to a 206

small dry N2 flow via a T-piece for a few hours. Note that the N2 flow was not delivered into 207

the vial but instead served a dry atmosphere at ~760 Torr. Cl2O was boiled at 5 oC and diffused 208

passively into the N2 flow. Cl2 was also removed because its boiling temperature is -34 oC 209

(Stull, 1947). The amount of N2O5 in ClONO2 was minimized because Cl2O was in excess. In 210

addition, the vapour pressure (a few mTorr) of N2O5 (Stull, 1947) is >100 times lower than that 211

of ClONO2 (~1 Torr) at around -76 oC (Schack and Lindahl, 1967; Ballard et al., 1988; 212

Anderson and Fahey, 1990); therefore, even if N2O5 was present in the gas phase, its amount 213

would be negligible compared to ClONO2. 214


10

2.1.3 ClONO2 detection 215

The ClONO2 vial was stored at -76 oC in the dark using a cryostat. A small dry N2 flow 216

(a few mL/min, F3) was delivered into the vial to elute gaseous ClONO2. The ClONO2 flow 217

was delivered through 1/8’’ FEP tubing in a stainless-steel injector into the centre of the aerosol 218

flow tube. The position of the injector could be adjusted to vary the interaction time of ClONO2 219

with aerosols in the flow tube. 220

The flow sampled from the flow tube (500 mL/min) was mixed with ~5 mL/min NO (100 221

ppmv in N2) and then delivered into a glass reactor heated to 130 oC. The initial NO mixing 222

ratio (in the absence of ClONO2) in the reactor was ~1000 ppbv (or ~1.8×1013 molecule cm-3). 223

The volume of the glass reactor (inner diameter: 2.0 cm; length: 10 cm) is ~30 cm3, 224

corresponding to an average residence time of ~2.6 s at 130 oC. ClONO2 was thermally 225

decomposed in the reactor to ClO and NO2 (R2, where M is the third molecule, e.g., N2), and 226

ClO was then titrated by NO in excess (R3): 227

ClONO2 + M → ClO +NO2 + M (R2) 228

ClO + NO → Cl + NO2 (R3) 229

Cl atoms produced in reaction (R3) further reacted with ClONO2 (R4), and the NO3 radicals 230

formed were titrated by NO (R5): 231

Cl + ClONO2 → Cl2 + NO3 (R4) 232

NO3 + NO → NO2 + NO2 (R5) 233

If the thermal dissociation of ClONO2 (R2) and the scavenging of ClO and NO3 radicals by 234

NO (R3, R4) all reach completion, the initial mixing ratio of ClONO2 is equal to the decrease 235

in the NO mixing ratios before and after introducing ClONO2 into the reactor (Anderson and 236

Fahey, 1990). 237

The lifetime of ClONO2 with respect to thermal dissociation (R2) at 130 oC was estimated 238

to be ~0.2 s at 160 Torr (Anderson and Fahey, 1990), and further increase in pressure to ~760 239


11

Torr would increase the decomposition rate and reduce its lifetime. The lifetime of ClONO2 240

with respect to reaction (R4) is not critical for our purpose although it enhances the overall 241

decay of ClONO2 in the reactor. The second order rate constants are 1.3×10-11 cm3 molecule-242

1 s-1 for the reaction of ClO with NO and 2.3×10-11 cm3 molecule-1 s-1 for the reaction of NO3 243

with NO at 130 oC (Burkholder et al., 2015), giving lifetimes of ~4×10-3 s for ClO with respect 244

to reaction (R3) and ~2×10-3 s for NO3 with respect to reaction (R5) in the presence of ~1000 245

ppbv NO in the reactor. To conclude, under our experimental conditions, the residence time of 246

the gas flow in the heated reactor was long enough for the completion of thermal dissociation 247

of ClONO2 (R2) and titrations of ClO and NO3 by NO (R3 and R5). 248

The flow exiting the reactor was sampled by a chemiluminescence-based NOx analyser 249

(Model 200E, Teledyne Instruments, USA), which has a sampling flow rate of 500 mL/min 250

(±10%). This instrument has two modes. In the first mode NO is measured by detecting the 251

chemiluminescence of exited NO2 (NO2*) produced by reacting NO with O3 in excess. The gas 252

flow can also be passed through a convertor cartridge filled with molybdenum (Mo) chips 253

heated to 315 oC and all the NO2 (and very likely also some of other NOy, e.g., HONO, HNO3) 254

is converted to NO; in this mode the total NO (initial NO and NO converted from NO2 etc.) is 255

measured and termed as NOx. The two modes are periodically switched, and the instrument 256

has a detection limit of 0.5 ppbv with a time resolution of 1 min. 257

The response of measured NO and NOx mixing ratios to the introduction of ClONO2 into 258

the AFT is displayed in Figure 2. Both the sheath flow and the flow in the AFT were set to 259

1500 mL/min (dry N2), and the injector was at 40 cm. The introduction of ClONO2 into the 260

AFT at ~20 min leads to the decrease of NO (solid curve in Figure 2a) from ~1100 ppbv to 261

~400 ppbv, and NO recovered to its initial level after stopping the ClONO2 flow at ~120 min. 262

The ClONO2 mixing ratio (solid curve in Figure 2b), derived from the change in the NO mixing 263

ratio, was very stable over 100 min. As expected, the introduction of ClONO2 into the system 264


12

led to the increase of the measured NOx mixing ratio (dashed curve in Figure 2a). Ideally the 265

increase in NOx mixing ratios due to the introduction of ClONO2 should be equal to the 266

ClONO2 mixing ratio. The nitrogen balance (dashed curve in Figure 2b), defined as the 267

difference in the ClONO2 mixing ratios (equal to the change in NO mixing ratios) and the 268

change of the NOx mixing ratios, is essentially zero within the experimental noise level. This 269

gives us further confidence in the purity of our synthesized ClONO2: under our current 270

detection scheme the change in the NOx mixing ratios will be twice of the N2O5 mixing ratio, 271

and therefore N2O5 contained in the ClONO2 flow as an impurity was negligible. This method 272

provides a simple and relatively selective method to quantify ClONO2, and could be used to 273

calibrate other ClONO2 detection methods (Anderson and Fahey, 1990). One previous study 274

used a similar method to detect ClONO2 in their experiments of ClONO2 uptake onto sulfuric 275

acid aerosol particles (Ball et al., 1998), with the only difference being that in their study NO 276

was detected by its absorption at 1845.5135 cm-1. Their reported γ(ClONO2) onto sulfuric acid 277

aerosol particles are in good agreement with those measured by other studies in which ClONO2 278

was measured using mass spectroscopy. This suggests that the indirect detection method of 279

ClONO2 utilized by Ball et al. (1998) and in this work can be used to investigate the uptake of 280

ClONO2 onto aerosol particles. 281

2.2 Coated-wall flow tube 282

The coated-wall flow tube, a Pyrex glass tube with an inner diameter of 30 mm, was used 283

to measure the uptake of ClONO2 onto fresh Pyrex glass. The inner wall was rinsed with diluted 284

NaOH solution and then by methanol and deionized water. A flow of 1500 mL/min, humidified 285

to the desired RH, was delivered into the top of the flow tube via a side arm. A small N2 flow 286

was used to elute the liquid ClONO2 sample, and the flow was then delivered through a 1/8’’ 287

Teflon tube in a stainless steel injector into the centre of the flow tube. The position of the 288

injector could be changed to vary the interaction time between ClONO2 and the inner wall of 289


13

the flow tube. At the bottom of the flow tube, a flow of 500 mL/min was sampled through 290

another side arm, mixed with ~5 mL/min NO (100 ppmv in N2), and then delivered into a glass 291

reactor heated to 130 oC. The flow exiting the heated glass reactor was then sampled into a 292

NOx analyser. The method used to detect ClONO2 is detailed in Section 2.1. The remaining 293

flow (~1000 mL/min) went through a RH sensor into the exhaust. 294

The linear flow velocity in the flow tube is 3.54 cm s-1 with a Reynolds number of 69, 295

suggesting that the flow is laminar. The length of the flow tube, defined as the distance between 296

the side arm through which the main flow was delivered into the flow tube and the other side 297

arm through which 500 mL/min was sampled from the flow tube into the NOx analyser, is 100 298

cm, giving a maximum residence time of ~30 s. The entrance length required to fully develop 299

the laminar flow and the mixing length required to fully mix ClONO2 with the main flow are 300

both less than 15 cm. The loss of ClONO2 onto the inner wall was measured using the middle 301

part (30-80 cm) of the flow tube. 302

2.3 Chemicals 303

NO (>99% purity) in a lecture bottle and the 100 ppmv (±1 ppmv) NO in N2 were supplied 304

by CK Special Gas (UK). Pure Cl2 (with a purity of >99.5%) in a lecture bottle and HgO 305

(yellow powder, with a purity of >99%) were provided by Sigma-Aldrich (UK). N2 and O2 306

were provided by BOC Industrial Gases (UK). P25 TiO2, with an anatase to rutile ratio of 3:1, 307

was supplied by Degussa-Hüls AG (Germany). SiO2 powders with a stated average particle 308

size (aggregate) of 200-300 nm were purchased from Sigma-Aldrich (UK). The BET surface 309

area is 8.3 m2 g-1 for TiO2 (Tang et al., 2014d) and ~201 m2 g-1 for SiO2 (Tang et al., 2014a). 310

3 Model description 311

The UKCA chemistry-climate model in its coupled stratosphere-troposphere 312

configuration, which combines both the tropospheric (O'Connor et al., 2014) and stratospheric 313

(Morgenstern et al., 2009) schemes, was used to simulate the effect of heterogeneous 314


14

hydrolysis of N2O5 (R1a) and ClONO2 (R1b) on TiO2. In this model the chemical cycles of Ox, 315

HOx and NOx, the oxidation of CO, ethane, propane and isoprene, chlorine and bromine 316

chemistry, and heterogeneous reactions on sulfuric acid and polar stratospheric clouds are all 317

included. The same approach used to investigate the effects of the eruption of Mt. Pinatubo on 318

stratospheric ozone (Telford et al., 2009) is adopted in this study. Telford et al. (2009) used the 319

UKCA model in a “nudged” configuration by constraining the dynamics of the model to 320

observations (Telford et al., 2008), and evaluated the difference of stratospheric ozone with 321

and without the additional sulfuric acid aerosols caused by the eruption of Mt. Pinatubo. 322

In our current study three simulations are used to assess the effects of TiO2 particle 323

injection into the stratosphere. All three simulations are started from a spun-up initial condition 324

and run from December 1990 to January 1993. In the base scenario (S1), an aerosol climatology 325

is used which represents the background loading of stratospheric sulphate aerosol. Alongside 326

S1 two further simulations were performed, one representing the eruption of Mt. Pinatubo in 327

June 1991 (S2) and a second (S3) in which the Mt. Pinatubo eruption is replaced with a single 328

injection of TiO2 particles on the same date. The simulations are set up so that the radiative 329

impacts at the surface are comparable between S2 and S3. Pope et al. (2012) have proposed 330

that 10 Tg of TiO2 aerosol particles with an assumed radius of 70 nm are required in order to 331

achieve the same solar radiation scattering effect as the eruption of Mt. Pinatubo. The total 332

surface area of TiO2 is calculated from the mass of TiO2 particles, using a density of 4.23 g 333

cm-3 and an assumed radius of 70 nm, and the global distribution of TiO2 is scaled to the sulfuric 334

acid aerosol distribution resulting from the eruption of Mt. Pinatubo. The sulfuric acid aerosol 335

surface area distribution was derived from the SPARC climatology (SPARC, 2006). 336

By running these three scenarios we are able to compare the relative impact of 337

stratospheric particle injection using TiO2 compared to sulphate. The benefit of using the Mt. 338

Pinatubo eruption as the sulphate injection scenario is that it provides a natural analogue to 339


15

proposed climate engineering schemes, and the chemical and dynamical effects of the eruption 340

have been well documented. Telford et al. (2009) have shown that UKCA accurately models 341

the chemical impacts of the Mt. Pinatubo eruption, and the ozone bias is smaller now compared 342

to Telford et al. (2009). 343

It should be noted that all simulations are nudged to the same observed meteorological 344

conditions, following Telford et al. (2008). In this way we do not take into account the 345

radiative/dynamical feedbacks from any ozone changes resulting from chemical reactions 346

occurring on stratospheric aerosols, allowing just the chemical effects of stratospheric particle 347

injection to be quantified. The results presented here expand on our previous study (Tang et al., 348

2014d) by including heterogeneous hydrolysis of both N2O5 (R1a) and ClONO2 (R1b) on TiO2. 349

An uptake coefficient of 1.5×10-3 is used for R1a (reaction with N2O5) on TiO2 particles as 350

determined by our previous measurement (Tang et al., 2014d). An uptake coefficient of 351

1.5×10-3 is used for R1b (heterogeneous hydrolysis of ClONO2), which is slightly larger than 352

the experimental value measured in our current work (i.e. ~1.2×10-3 as shown in Section 4) and 353

represents an upper limit to the measured value given the experimental uncertainty. We use 354

this value to determine the maximum impact TiO2 injection would have on stratospheric 355

chemistry. 356

4 Results& Discussion 357

4.1 Uptake of ClONO2 onto Pyrex glass 358

The uptake of ClONO2 onto fresh Pyrex glass wall was determined by measuring the 359

ClONO2 concentrations at five different injection positions. The loss of ClONO2 in the coated-360

wall flow tube, under the assumption of pseudo first order kinetics, can be described by the Eq. 361

(1): 362

[𝐶𝑙𝑂𝑁𝑂2]𝑡 = [𝐶𝑙𝑂𝑁𝑂2]0 ∙ exp(−𝑘𝑤 ∙ 𝑡) (1) 363


16

where [ClONO2]t and [ClONO2]0 are the measured ClONO2 concentrations at the reaction time 364

of t and 0, respectively, and kw is the wall loss rate (s-1). Two typical datasets of measured 365

[ClONO2] at five different injector positions are displayed in Figure 3, suggesting that ClONO2 366

indeed follows the exponential decays, and the slopes of the exponential decays are equal to 367

kw. The effective (or experimental) uptake coefficient of ClONO2, γeff, onto the Pyrex wall, can 368

then be calculated from kw, using Eq. (2) (Howard, 1979; Wagner et al., 2008): 369

𝛾𝑒𝑓𝑓 =𝑘𝑤∙𝑑𝑡𝑢𝑏𝑒

𝑐(𝐶𝑙𝑂𝑁𝑂2) (2) 370

where dtube is the inner diameter of the flow tube (3.0 cm) and c(ClONO2) is the average 371

molecular speed of ClONO2 (25 360 cm s-1). Depletion of ClONO2 close to the wall is caused 372

by the uptake of ClONO2 onto the wall, and thus the effective uptake coefficient is smaller than 373

the true one. This effect can be corrected (Tang et al., 2014b), and the true uptake coefficients, 374

γ, are reported in Table 1. 375

The uptake coefficients of ClONO2 onto Pyrex glass, as summarized in Table 1, increases 376

from ~5×10-6 at 0% RH to ~1.6×10-5 at 24% RH by a factor of ~3. Uptake coefficients at higher 377

RH were not determined because the uptake coefficients determined at 24% RH (~1.6×10-5) 378

are very close to the upper limit (~2.3×10-5) which can be measured in this study using the 379

coated-wall flow tube technique due to the gas phase diffusion limit. The RH dependence of 380

γ(ClONO2) for Pyrex glass is further discussed in Section 4.4 together with these reported by 381

Molina et al. (1997) and our measurements on SiO2 and TiO2 aerosol particles. 382

4.2 Reaction of ClONO2 with SiO2 and TiO2 particles 383

The uptake of ClONO2 onto airborne SiO2 and TiO2 particles were investigated using an 384

atmospheric pressure aerosol flow tube, in which reactions with the aerosol particles and the 385

wall both contribute to the loss of ClONO2, as shown in Eq. (3): 386

[𝐶𝑙𝑂𝑁𝑂2]𝑡 = [𝐶𝑙𝑂𝑁𝑂2]0 ∙ exp[−(𝑘𝑤 + 𝑘𝑎) ∙ 𝑡] (3) 387


17

where [ClONO2]t and [ClONO2]0 are the measured ClONO2 mixing ratios at the reaction times 388

of t and 0 s, and kw and ka are the loss rates (s-1) of ClONO2 onto the inner wall of the flow tube 389

and the surface of aerosol particles, respectively. In a typical uptake measurement, the aerosol 390

flow was delivered through a filter, and [ClONO2] was measured at five different injector 391

positions to determine the wall loss rate (kw). The filter was then bypassed to deliver aerosol 392

particles into the flow tube, and the total ClONO2 loss rate (kw + ka) in the flow tube was 393

determined. After that, the aerosol flow was passed through the filter to measure kw again. The 394

variation of kw determined before and after introducing particles into the flow tube was within 395

the experimental uncertainty of kw, ensuring that the reactivity of the wall towards ClONO2 396

remained constant during the uptake measurement. 397

The difference between the ClONO2 loss rates without and with aerosol particles in the 398

flow tube, is equal to the loss rate due to the reaction with surface of aerosol particle (ka). The 399

effective uptake coefficient of ClONO2 onto aerosol particles, γeff, is related to ka by Eq. (4) 400

(Crowley et al., 2010): 401

𝑘𝑎 = 0.25 ∙ 𝛾𝑒𝑓𝑓 ∙ 𝑐(𝐶𝑙𝑂𝑁𝑂2) ∙ 𝑆𝑎 (4) 402

where Sa is the aerosol surface area concentration which can be derived from size-resolved 403

number concentrations (as shown in Figure S1) measured by the SMPS. Uptake of ClONO2 404

onto aerosol particles also leads to the depletion of ClONO2 near the particle surface and so the 405

effective uptake coefficient is smaller than the true uptake coefficient. This effect, which can 406

be corrected using the method described elsewhere (Tang et al., 2014b), is only a few percent 407

in this study as the particle diameters are <1 μm and the uptake coefficient is relatively small 408

(~1×10-3). 409

Two typical decays of ClONO2 in the aerosol flow tube without and with SiO2/TiO2 410

aerosol particles in the flow tube are shown in Figure 4. For a majority of experiments, efforts 411

were made to generate enough aerosol particles so that ka+kw was significantly different to kw. 412


18

It is evident from Figure 4 that the loss of ClONO2 is significantly faster with TiO2/SiO2 413

particles in the flow tube than without aerosols. We acknowledge that the measured ka and 414

therefore our reported γ in this study have quite large uncertainties. This is because the uptake 415

coefficients of ClONO2 are very small and the surface area of the wall is ~1000 larger than that 416

of aerosol particles. This is the first time that heterogeneous reactions of ClONO2 with airborne 417

mineral particles have been investigated. 418

The uptake coefficients of ClONO2 are ~1.2×10-3 for TiO2 particles, and no difference in 419

γ(ClONO2) at two different RH (7% and 33%) is found. The heterogeneous reaction of ClONO2 420

with SiO2 particles were studied at four different RH, with γ(ClONO2) increasing from ~2×10-4 421

at 7% RH to ~5×10-4 at 35% RH, reaching a value of ~6×10-4 at 59% RH. The uptake 422

coefficients of ClONO2 are summarized in Table 2 for SiO2 and TiO2 aerosol particles, together 423

with key experimental conditions. In a few measurements in which the SiO2 aerosol 424

concentrations were relatively low, the total ClONO2 loss rate (kw + ka) was not different from 425

its wall loss rate (kw) within the experimental uncertainty. In this case, only the upper limit of 426

ka (and thus γ) can be estimated, which is reported here as the standard deviation of kw. The 427

first three of the four uptake coefficients at (17±2)% RH for SiO2 aerosol particles, tabulated 428

in Table 2, fall into this category. γ(ClONO2) on SiO2 aerosol particles is around two orders of 429

magnitude larger than that on Pyrex glass. One explanation for such a large difference is that 430

SiO2 particles used in our work are porous (Tang et al., 2014a) and therefore the surface area 431

which is actually available for the ClONO2 uptake is much larger than that calculated using the 432

mobility diameters. In our previous study (Tang et al., 2014a) we have found that for SiO2 433

particles, γ(N2O5) calculated using the mobility diameter based surface area are a factor of 40 434

larger than those calculated using the BET surface area. Another reason is that the composition 435

of SiO2 is different from Pyrex. 436


19

4.3 Effects of RH 437

The RH dependence of γ(ClONO2) for Pyrex glass is plotted in Figure 5 and exhibits a 438

positive dependence on RH, with γ(ClONO2) increased by a factor of ~3 when RH increases 439

from 0% to 24%. Previous studies (Hanson and Ravishankara, 1991; Hanson and Ravishankara, 440

1994; Zhang et al., 1994; Hanson, 1998) have shown that γ(ClONO2) for aqueous H2SO4 441

solution strongly depends on water content in the solution and it decreases from ~0.1 for 40% 442

H2SO4 to ~1×10-4 for 75% H2SO4 at 200-200 K, by a factor of ~1000. It is suggested that the 443

heterogeneous uptake of ClONO2 by aqueous H2SO4 solution proceeds via direct and acid-444

catalysed hydrolysis (Robinson et al., 1997; Shi et al., 2001; Ammann et al., 2013). One may 445

expect that γ(ClONO2) for Pyrex glass will increase with RH. This is also supported by the 446

water adsorption isotherm on Pyrex glass particles (Chikazawa et al., 1984), showing that the 447

amount of adsorbed water on Pyrex surface displays a substantial increase at 20% RH 448

compared to that at 0% RH. However, the results reported by Chikazawa et al. (1984) are 449

presented graphically and thus impede us from a more quantitative discussion on the effect of 450

RH and surface-adsorbed water on uptake of ClONO2 by Pyrex surface. 451

One can then expect that γ(ClONO2) may also increase with RH for the reaction with 452

SiO2 and TiO2 aerosol particles, since the amount of water adsorbed on these two types of 453

particles also increase with RH (Goodman et al., 2001). Inspection of the data listed in Table 2 454

reveals that γ(ClONO2) for SiO2 particle increases from ~2×10-4 at 7% RH to ~6×10-4 at 59% 455

RH, and this is consistent with the large increase of adsorbed water on SiO2 surface, from 456

around half a monolayer at ~7% RH to two monolayers at 60% RH (Goodman et al., 2001), as 457

shown in Figure S2. The uptake coefficients of ClONO2 were measured to be ~1.2×10-3 for 458

TiO2 at 7% and 33% RH, with no significant difference found at these two different RH. We 459

expect that further increase in RH will lead to larger γ(ClONO2) for TiO2, and future studies at 460

higher RH are needed to better understand the RH effects. 461


20

At similar RH (7% and 33%), γ(ClONO2) for TiO2 are significantly larger than those for 462

SiO2. This may be explained by the larger amount of adsorbed water on TiO2 at low and 463

medium RH compared to SiO2 as shown in Figure S2. It is interesting to note that the uptake 464

of N2O5 shows different behaviour, i.e. γ(N2O5) for SiO2 (Tang et al., 2014a) are significantly 465

larger than that for TiO2 at similar RH. This indicates that a different mechanism controls N2O5 466

uptake by mineral surfaces (Seisel et al., 2005; Tang et al., 2014a; Tang et al., 2014c): at low 467

RH the reaction with surface OH groups is the major pathway for the uptake of N2O5 onto 468

SiO2/TiO2 surface, instead of direct hydrolysis by surface-adsorbed water. 469

4.4 Comparison with previous work 470

We find that in the absence of HCl, γ(ClONO2) is around 1.2×10-3 for TiO2 aerosol 471

particles and <1×10-3 for SiO2 aerosol particles at room temperature. Using the coated-wall 472

flow tube technique, Molina et al. (1997) investigated the uptake of ClONO2 onto the inner 473

wall of an Al2O3 tube, α-Al2O3 particles, and the inner wall of a Pyrex glass tube, in the 474

presence of (1-10)×10-6 Torr HCl at 200-220 K. Uptake coefficients of ~0.02 were reported for 475

all the three types of surface (including Pyrex glass), over a factor of 1000 larger than 476

γ(ClONO2) for Pyrex glass determined in our present work. The large difference in γ(ClONO2) 477

reported by the two studies is likely due to the co-presence of HCl (1×10-6-1×10-5 Torr) in the 478

experiments of Molina et al. (1997), while no HCl was present in our work. Heterogeneous 479

reactions of ClONO2 proceed via direct and acid-catalysed hydrolysis (Robinson et al., 1997; 480

Shi et al., 2001; Ammann et al., 2013), and numerous previous studies have confirmed that the 481

presence of HCl in the gas phase (and thus partitioning into or adsorption onto the condensed 482

phases) promotes the uptake of ClONO2 by H2SO4 solution, ice, and nitric acid trihydrate 483

(NAT), as summarized by Crowley et al. (2010), Sander et al. (2011) and Ammann et al. (2013). 484

Temperature may also play a role since measurements were carried out at 200-220 K by Molina 485

et al. (1997) and at ~296 K in our study. 486


21

Considering the importance of HCl in the ClONO2 uptake and its abundance in the 487

stratosphere, it will be important to systematically measure γ(ClONO2) for SiO2/TiO2 in the 488

presence of HCl over a broad HCl concentration and temperature range relevant for lower 489

stratosphere. 490

5 Implication for stratospheric particle injection 491

Injection of TiO2 into the stratosphere will provide additional surface area for the 492

heterogeneous reactions of N2O5 (R1a) and ClONO2 (R1b, R1c). There are several important 493

types of particles naturally present in the stratosphere (Solomon et al., 1999), including sulfuric 494

acid, ice, and nitric acid trihydrate (NAT), and their interaction with ClONO2 has been well 495

characterised (Crowley et al., 2010; Ammann et al., 2013; Burkholder et al., 2015). Comparing 496

γ(ClONO2) for TiO2 particles with these other stratospherically relevant surfaces can provide 497

a first order estimate of their relative importance. 498

The uptake of ClONO2 on H2SO4 acid particles is strongly influenced by temperature and 499

the water content in the particles (Shi et al., 2001; Ammann et al., 2013; Burkholder et al., 500

2015): γ(ClONO2) are <2×10-3 for 65wt% H2SO4 particles and <2×10-4 for 75wt% H2SO4 501

particles. The global distribution of γ(ClONO2) calculated for sulfuric acid particles in the 502

stratosphere is shown in the supporting information (Figure S3), suggesting that γ(ClONO2) is 503

lower on TiO2 particles than on sulfuric acid particles in the lower stratosphere. The uptake 504

coefficient of ClONO2 for water ice shows a negative dependence on temperature, with 505

γ(ClONO2) of ~0.1 at ~200 K (Crowley et al., 2010; Burkholder et al., 2015), around a factor 506

of 100 larger than that for TiO2 particles at room temperature. γ(ClONO2) for water-rich nitric 507

acid trihydrate (NAT), another important component for polar stratospheric clouds, increases 508

strongly with temperature, with γ(ClONO2) of 3.0×10-3 at 200 K, 6.0×10-3 at 210 K, and 509

1.14×10-2 at 220 K (Crowley et al., 2010). 510


22

While the background burden of stratospheric aerosol is low, volcanic eruptions and 511

deliberate stratospheric particle injection for climate engineering purposes have the potential 512

to significantly increase the available surfaces for heterogeneous reactions. In our current work, 513

three simulations were performed, one representing a low background loading of stratospheric 514

sulphate (<1 Tg) aerosols (S1), a second representing the eruption of Mt. Pinatubo (S2) and a 515

third representing an instantaneous injection of 10 Tg of TiO2 (S3). SiO2 particle injection is 516

not considered in our modelling study because the refractive index of SiO2 is significantly 517

smaller than TiO2 (Pope et al., 2012). Two heterogeneous reactions on TiO2 particles, i.e. 518

heterogeneous hydrolysis of N2O5 (R1a) and ClONO2 (R1b), were included in the simulation: 519

a value of 1.5×10-3 was used for γ(N2O5), as measured in our previous work (Tang et al., 2014d), 520

and γ(ClONO2) was also set to 1.5×10-3, based on the measurement reported in our current 521

study. All three simulations were nudged to observed meteorology from December 1990 to 522

January 1993. By comparing the TiO2 injection (S3) with the Mt. Pinatubo eruption (S2) we 523

are able to quantify the relative impacts of TiO2 and sulphuric acid injection on stratospheric 524

chemistry. Results in this section are presented as annual means for the year 1992. 525

Similar to our previous study (Tang et al., 2014d), we have found that injection of TiO2 526

(S3) has a much reduced impact on stratospheric N2O5 concentrations than the eruption of Mt. 527

Pinatubo (S2). N2O5 mixing ratios are significantly reduced in S2 compared to S1 from 10-30 528

km, with concentrations reduced by >80% throughout most of this region. For comparison, 529

after TiO2 injection (S3) N2O5 concentrations are reduced over a much smaller altitude range 530

(15-25 km) and to a lesser degree, with ~20% reductions in the tropics and up to 60% reductions 531

in the high latitudes. The relative effects of TiO2 injection compared to sulphate injection on 532

N2O5 mixing ratios is calculated as the difference between S3 and S2. As shown in Figure 6, 533

throughout most of the stratosphere N2O5 mixing ratios remain higher under S3 than S2. 534


23

Under both particle injection scenarios (S2 and S3), stratospheric ClOx mixing ratios are 535

increased compared to S1 due to the activation of ClONO2 through heterogeneous reactions. 536

However, Figure 7 suggests that ClOx mixing ratios are up to 40% lower in the tropical lower 537

stratosphere following the injection of TiO2 aerosols compared to sulphate. This is driven in 538

part by the lower surface area density of TiO2 compared to sulphate, but also due to the 539

difference in uptake coefficients. The uptake coefficient of ClONO2 onto sulphate is 540

temperature dependent, and our measurements suggest that the uptake coefficient onto TiO2 is 541

smaller than that for sulphate below ~215 K. Throughout much of the tropical lower 542

stratosphere where maximum aerosol surface area density is found in both S2 and S3, 543

temperatures are below ~220 K and therefore the uptake coefficient is lower for TiO2 than 544

sulphate (as shown by Figure S3 in the supporting information), leading to reduced chlorine 545

activation. Previous studies have investigated the influence of temperature on the 546

heterogeneous reactions of mineral particles with a few other trace gases, including HCOOH 547

(Wu et al., 2012), H2O2 (Romanias et al., 2012) and OH radicals (Bedjanian et al., 2013), and 548

found that the measured uptake coefficients varied only by a factor of 2-3 or less across a wide 549

temperature range. Therefore, the temperature effect is not expected to be very significant in 550

our current work, although further studies are required to quantify and reduce these 551

uncertainties. 552

The relative difference in ozone mixing ratios following TiO2 injection (S3) compared 553

with the eruption of Mt. Pinatubo (S2) is shown in Figure 8. Ozone mixing ratios in the lower 554

stratosphere decrease as a result of both TiO2 and sulphate injection, with largest decreases 555

seen at high latitudes. In terms of annual means, the magnitude of this ozone response is 556

comparable between the two simulations, with a maximum of ~3% in the tropics and ~7% at 557

high latitudes. In contract, ozone mixing ratios at the altitude of 25 km increase following the 558

eruption of Mt. Pinatubo (S2), but show no significant change upon TiO2 injection (S3). This 559


24

is consistent with the much faster uptake of N2O5 onto sulphate aerosols and the resultant 560

stratospheric NOx loss and decreases in the rates of catalytic ozone destruction at these altitudes. 561

The results presented here indicate that there is little difference in stratospheric ozone 562

concentrations between injection of TiO2 and sulphate aerosols when R1a and R1b are 563

considered on TiO2. While TiO2 injection (S3) leads to less ClOx activation and ozone 564

destruction in the lowermost stratosphere, the reduced depletion of N2O5 and NOx in the middle 565

stratosphere leads to decreased ozone mixing ratios compared to sulphate injection (S2). The 566

total column ozone differences between S3 and S2 are within ±2.5%, indicating that there is 567

no significant difference in vertically integrated ozone abundancies and solar UV amounts 568

reaching the surface. However, more work is required to establish additional kinetic data for 569

heterogeneous reactions of TiO2. 570

6 Conclusion 571

Minerals with high refractive indices, such as TiO2, have been proposed as possible 572

materials used for stratospheric particle injection for climate engineering (Pope et al., 2012). 573

However, kinetic data of their heterogeneous reactions with important reactive trace gases (e.g., 574

N2O5 and ClONO2) in the stratosphere are lacking, impeding us from a reliable assessment of 575

the impacts of mineral particle injection on stratospheric ozone in particular and stratosphere 576

chemistry in general. In our current work, using an atmospheric pressure aerosol flow tube, we 577

have investigated the heterogeneous reaction of ClONO2 with TiO2 and SiO2 aerosol particles 578

at room temperature and at different RH. The uptake coefficient, γ(ClONO2), was ~1.2×10-3 at 579

7% and 33% RH for TiO2 particles, with no significant difference observed at these two RH; 580

for SiO2 particles, γ(ClONO2) increases from ~2×10-4 at 7% RH to ~6×10-4 at 59%, showing a 581

positive dependence on RH. Therefore, it can be concluded that under similar conditions for 582

the RH range covered in this work, TiO2 shows higher heterogeneous reactivity than SiO2 583

towards ClONO2. Compared to sulfuric acid particles in the lower stratosphere, the 584


25

heterogeneous reactivity towards ClONO2 is lower for TiO2 particles. In addition, the 585

heterogeneous uptake of ClONO2 by Pyrex glass was also studied, with γ(ClONO2) increasing 586

from ~4.5×10-6 at 0% RH to ~1.6×10-5 at 24% RH. 587

Using the UKCA chemistry-climate model with nudged meteorology, we have 588

constructed a scenario to assess the impact of TiO2 particle injection on stratospheric chemistry. 589

In this scenario TiO2 aerosol particles are distributed in the stratosphere in such a way that TiO2 590

particle injection is assumed to produce a radiative effect similar to that of Mt. Pinatubo 591

eruption, following Pope et al. (2012). Heterogeneous reactions of N2O5 and ClONO2 with 592

TiO2 aerosol particles, both with an uptake coefficient of 1.5×10-3 based on our previous (Tang 593

et al., 2014d) and current laboratory experiments, were included in the simulation. It is found 594

that compared to the eruption of Mt. Pinatubo, the TiO2 injection has a much smaller impact 595

on N2O5 in the stratosphere, although significant reduction (20-60% compared to the 596

background scenario without additional particle injection) in stratospheric N2O5 also occurs. 597

Compared to the background scenario, both TiO2 injection and the Mt. Pinatubo eruption 598

scenarios lead to increased stratospheric ClOx mixing ratios, and the ClOx mixing ratios are 599

lower for the TiO2 injection than the Mt. Pinatubo eruption. Both TiO2 injection and the Mt. 600

Pinatubo eruption results in significant ozone depletion in the lower stratosphere, with largest 601

decreases occurring at high latitudes. In comparison with Mt. Pinatubo eruption, TiO2 injection 602

causes less ClOx activation and less ozone destruction in the lowermost stratosphere, while the 603

reduced depletion of N2O5 and NOx in the middle stratosphere results in decreased ozone levels. 604

Overall, our simulation results suggest that there is no significant difference (within ±2.5%) in 605

the vertically integrated ozone abundancies between TiO2 injection and Mt. Pinatubo eruption. 606

It should be emphasized that heterogeneous chemistry of TiO2 included in our current 607

modelling study is not complete. One example is the heterogeneous reaction of ClONO2 with 608

HCl (R1c) on/in the particles. An uptake coefficient of 0.02 was reported for the heterogeneous 609


26

reaction of ClONO2 with HCl on Al2O3 particles (Molina et al., 1997), and it is reasonable to 610

assume that this reaction may also be quite fast on TiO2 particles. The heterogeneous reaction 611

of ClONO2 with HCl on TiO2 particles, with an uptake coefficient assumed to be the same as 612

that on Al2O3 surface (i.e. 0.02) as reported by Molina et al. (1997), has been included in further 613

simulations, and the results will be reported and discussed in a following paper. Other reactions, 614

including the heterogeneous reaction of HOCl (Molina et al., 1996; Solomon, 1999) and a range 615

of heterogeneous photochemical reactions (Chen et al., 2012; George et al., 2015), may also be 616

important and thus deserve further laboratory and modeling investigation. 617

Our nudged modeling simulations, designed to focus on chemistry effects, do not take 618

into account feedbacks between radiative effects, atmospheric dynamics, and chemistry. 619

Several recent studies have assessed the impact of high latitude stratospheric ozone depletion 620

using the UKCA model (Braesicke et al., 2013; Keeble et al., 2014) and have shown that 621

interactive feedbacks can affect stratospheric temperatures, the strength of the Brewer-Dobson 622

circulation, the longevity of polar vortices and surface climate. By nudging the model to 623

observed meteorology during the Mt. Pinatubo eruption these feedbacks are implicitly included 624

in the sulphate injection scenario. However, while we have chosen a TiO2 loading to give the 625

same surface radiative response as the Mt. Pinatubo eruption, the stratospheric radiative 626

impacts may differ. In order to fully understand the true impact of stratospheric particle 627

injection, both the radiative and chemical effects, and the coupling between these responses, 628

need to be explored further. In addition, before any climate engineering schemes could be 629

considered, much consideration is absolutely obligatory, including, but not limited to, technical, 630

socioeconomic, political, environmental, and ethical feasibilities. 631

Acknowledgement 632

Financial support provided by EPSRC Grant EP/I01473X/1 and the Isaac Newton Trust 633

(Trinity College, University of Cambridge, U.K.) is acknowledged. We thank NCAS-CMS for 634


27

modelling support. Model integrations have been performed using the ARCHER UK National 635

Supercomputing Service. We acknowledge the ERC for support through the ACCI project 636

(project number: 267760). 637


28

Reference 638

Ammann, M., Cox, R. A., Crowley, J. N., Jenkin, M. E., Mellouki, A., Rossi, M. J., Troe, J., 639

and Wallington, T. J.: Evaluated kinetic and photochemical data for atmospheric chemistry: 640

Volume VI - heterogeneous reactions with liquid substrates, Atmos. Chem. Phys., 12, 8045-641

8228, 2013. 642

Anderson, L. C., and Fahey, D. W.: Studies with ClONO2: Thermal-dissociation rate and 643

catalytic conversion to NO using an NO/O3 chemi-luminescence detector, J. Phys. Chem., 94, 644

644-652, 1990. 645

Ball, S. M., Fried, A., Henry, B. E., and Mozurkewich, M.: The hydrolysis of ClONO2 on 646

sub-micron liquid sulfuric acid aerosol, Geophys. Res. Lett., 25, 3339-3342, 1998. 647

Ballard, J., Johnston, W. B., Gunson, M. R., and Wassell, P. T.: Absolute absorption 648

coefficients of ClONO2 infrared bands at stratospheric temperatures, J. Geophys. Res.-649

Atmos, 93, 1659-1665, 1988. 650

Bannan, T. J., Booth, A. M., Bacak, A., Muller, J. B. A., Leather, K. E., Le Breton, M., Jones, 651

B., Young, D., Coe, H., Allan, J., Visser, S., Slowik, J. G., Furger, M., Prévôt, A. S. H., Lee, 652

J., Dunmore, R. E., Hopkins, J. R., Hamilton, J. F., Lewis, A. C., Whalley, L. K., Sharp, T., 653

Stone, D., Heard, D. E., Fleming, Z. L., Leigh, R., Shallcross, D. E., and Percival, C. J.: The 654

first UK measurements of nitryl chloride using a chemical ionization mass spectrometer in 655

central London in the summer of 2012, and an investigation of the role of Cl atom oxidation, 656

J. Geophys. Res.-Atmos, 120, 5638-5657, 2015. 657

Bedjanian, Y., Romanias, M. N., and El Zein, A.: Interaction of OH Radicals with Arizona 658

Test Dust: Uptake and Products, J. Phys. Chem. A, 117, 393-400, 2013. 659

Braesicke, P., Keeble, J., Yang, X., Stiller, G., Kellmann, S., Abraham, N. L., Archibald, A., 660

Telford, P., and Pyle, J. A.: Circulation anomalies in the Southern Hemisphere and ozone 661

changes, Atmos. Chem. Phys., 13, 10677-10688, 2013. 662

Burkholder, J. B., Sander, S. P., Abbatt, J. P. D., Barker, J. R., Huie, R. E., Kolb, C. E., 663

Kurylo, M. J., Orkin, V. L., Wilmouth, D. M., and Wine, P. H.: Chemical Kinetics and 664

Photochemical Data for Use in Atmospheric Studies, Evaluation No. 18, JPL Publication 15-665

10, Jet Propulsion Lab., Pasadena, CA, 2015. 666

Chen, H. H., Nanayakkara, C. E., and Grassian, V. H.: Titanium Dioxide Photocatalysis in 667

Atmospheric Chemistry, Chem. Rev., 112, 5919-5948, 2012. 668

Chikazawa, M., Kanazawa, T., and Yamaguchi, T.: The Role of Adsorbed Water on 669

Adhesion Force of Powder Particles, KONA, 2, 54-61, 1984. 670

Crowley, J. N., Ammann, M., Cox, R. A., Hynes, R. G., Jenkin, M. E., Mellouki, A., Rossi, 671

M. J., Troe, J., and Wallington, T. J.: Evaluated Kinetic and Photochemical Data for 672

Atmospheric Chemistry: Volume V - Heterogeneous Reactions on Solid Substrates, Atmos. 673

Chem. Phys., 10, 9059-9223, 2010. 674

Crutzen, P. J.: Albedo enhancement by stratospheric sulfur injections: A contribution to 675

resolve a policy dilemma?, Clim. Change, 77, 211-219, 2006. 676

Danilin, M. Y., Shia, R. L., Ko, M. K. W., Weisenstein, D. K., Sze, N. D., Lamb, J. J., Smith, 677

T. W., Lohn, P. D., and Prather, M. J.: Global stratospheric effects of the alumina emissions 678

by solid-fueled rocket motors, J. Geophys. Res.-Atmos, 106, 12727-12738, 2001. 679

Davidson, J. A., Cantrell, C. A., Shetter, R. E., McDaniel, A. H., and Calvert, J. G.: Absolute 680

infrared absorption cross sections for ClONO2 at 296 and 223 K, J. Geophys. Res.-Atmos., 681

92, 10921-10925, 1987. 682

Dee, D. P., Uppala, S. M., Simmons, A. J., Berrisford, P., Poli, P., Kobayashi, S., Andrae, U., 683

Balmaseda, M. A., Balsamo, G., Bauer, P., Bechtold, P., Beljaars, A. C. M., van de Berg, L., 684

Bidlot, J., Bormann, N., Delsol, C., Dragani, R., Fuentes, M., Geer, A. J., Haimberger, L., 685

Healy, S. B., Hersbach, H., Holm, E. V., Isaksen, L., Kallberg, P., Kohler, M., Matricardi, 686


29

M., McNally, A. P., Monge-Sanz, B. M., Morcrette, J. J., Park, B. K., Peubey, C., de Rosnay, 687

P., Tavolato, C., Thepaut, J. N., and Vitart, F.: The ERA-Interim reanalysis: configuration 688

and performance of the data assimilation system, Q. J. R. Meteorol. Soc., 137, 553-597, 2011. 689

Deiber, G., George, C., Le Calvé, S., Schweitzer, F., and Mirabel, P.: Uptake study of 690

ClONO2 and BrONO2 by Halide containing droplets, Atmos. Chem. Phys., 4, 1291-1299, 691

2004. 692

Dutton, E. G., and Christy, J. R.: Solar radiative forcing at selected locations and evidence for 693

global lower tropospheric cooling following the eruptions of El Chichón and Pinatubo, 694

Geophys. Res. Lett., 19, 2313-2316, 1992. 695

Fahey, D. W., Eubank, C. S., Hubler, G., and Fehsenfeld, F. C.: A Calibrated Source of N2O5, 696

Atmos. Environ., 19, 1883-1890, 1985. 697

Fahey, D. W., Kawa, S. R., Woodbridge, E. L., Tin, P., Wilson, J. C., Jonsson, H. H., Dye, J. 698

E., Baumgardner, D., Borrmann, S., Toohey, D. W., Avallone, L. M., Proffitt, M. H., 699

Margitan, J., Loewenstein, M., Podolske, J. R., Salawitch, R. J., Wofsy, S. C., Ko, M. K. W., 700

Anderson, D. E., Schoeberl, M. R., and Chan, K. R.: In situ measuremnet constraining the 701

role of sulfate aerosol in midelatitude ozone depletion, Nature, 363, 509-514, 1993. 702

Fernandez, M. A., Hynes, R. G., and Cox, R. A.: Kinetics of ClONO2 reactive uptake on ice 703

surfaces at temperatures of the upper troposphere, J. Phys. Chem. A, 109, 9986-9996, 2005. 704

Ferraro, A. J., Highwood, E. J., and Charlton-Perez, A. J.: Stratospheric heating by potential 705

geoengineering aerosols, Geophys. Res. Lett., 38, L24706, 10.1029/2011GL049761, 2011. 706

Finlayson-Pitts, B. J., Ezell, M. J., and Pitts, J. N.: Formation of Chemically Active Chlorine 707

Compounds by Reactions of Atmospheric NaCl Particles with Gaseous N2O5 and ClONO2, 708

Nature, 337, 241-244, 1989. 709

George, C., Ammann, M., D’Anna, B., Donaldson, D. J., and Nizkorodov, S. A.: 710

Heterogeneous Photochemistry in the Atmosphere, Chem. Rev., 115, 4218-4258, 2015. 711

Ginoux, P., Prospero, J. M., Gill, T. E., Hsu, N. C., and Zhao, M.: Global-scale Attribution of 712

Anthropogenic and Natural Dust Sources and Their Emission Rates Based on MODIS Deep 713

Blue Aerosol Products, Rev. Geophys., 50, RG3005, doi: 3010.1029/2012RG000388, 2012. 714

Goodman, A. L., Bernard, E. T., and Grassian, V. H.: Spectroscopic Study of Nitric Acid and 715

Water Adsorption on Oxide Particles: Enhanced Nitric Acid Uptake Kinetics in the Presence 716

of Adsorbed Water, J. Phys. Chem. A, 105, 6443-6457, 2001. 717

Guo, S., Bluth, G. J. S., Rose, W. I., Watson, I. M., and Prata, A. J.: Re-evaluation of SO2 718

release of the 15 June 1991 Pinatubo eruption using ultraviolet and infrared satellite sensors, 719

Geochem. Geophys. Geosyst., 5, Q04001, doi: 04010.01029/02003GC000654, 2004. 720

Hanson, D. R., and Ravishankara, A. R.: The reaction probabilities of ClONO2 and N2O5 on 721

polar stratospheric cloud materials, J. Geophys. Res.-Atmos., 96, 5081-5090, 1991. 722

Hanson, D. R., and Ravishankara, A. R.: Reactive uptake of ClNO2 onto sulfuric-acid due to 723

reaction with HCl and H2O, J. Phys. Chem., 98, 5728-5735, 1994. 724

Hanson, D. R., and Lovejoy, E. R.: The Reaction of ClONO2 with Submicrometer Sulfuric-725

Acid Aerosol, Science, 267, 1326-1328, 1995. 726

Hanson, D. R.: Reaction of ClONO2 with H2O and HCl in sulfuric acid and 727

HNO3/H2SO4/H2O mixtures, J. Phys. Chem. A, 102, 4794-4807, 1998. 728

Hinds, W. C.: Aerosol techniques: properties, behavior, and measurement if airborne 729

particles, John Wiley& Sons. Inc., New York, 1996. 730

Howard, C. J.: Kinetic Measurements Using Flow Tubes, J. Phys. Chem., 83, 3-9, 1979. 731

Jackman, C. H., Considine, D. B., and Fleming, E. L.: A global modeling study of solid 732

rocket aluminum oxide emission effects on stratospheric ozone, Geophys. Res. Lett., 25, 907-733

910, 1998. 734


30

Jones, A. C., Haywood, J. M., and Jones, A.: Climatic impacts of stratospheric 735

geoengineering with sulfate, black carbon and titania injection, Atmos. Chem. Phys., 16, 736

2843-2862, 2016. 737

Kebede, M. A., Scharko, N. K., Appelt, L. E., and Raff, J. D.: Formation of Nitrous Acid 738

during Ammonia Photooxidation on TiO2 under Atmospherically Relevant Conditions, J. 739

Phys. Chem. Lett., 4, 2618-2623, 2013. 740

Keeble, J., Braesicke, P., Abraham, N. L., Roscoe, H. K., and Pyle, J. A.: The impact of polar 741

stratospheric ozone loss on Southern Hemisphere stratospheric circulation and climate, 742

Atmos. Chem. Phys., 14, 13705-13717, 2014. 743

Kravitz, B., Robock, A., Forster, P. M., Haywood, J. M., Lawrence, M. G., and Schmidt, H.: 744

An overview of the Geoengineering Model Intercomparison Project (GeoMIP), J. Geophys. 745

Res.-Atmos., 118, 13103-13107, 2013. 746

Lawler, M. J., Sander, R., Carpenter, L. J., Lee, J. D., von Glasow, R., Sommariva, R., and 747

Saltzman, E. S.: HOCl and Cl2 observations in marine air, Atmos. Chem. Phys., 11, 7617-748

7628, 2011. 749

Lee, S. H., Leard, D. C., Zhang, R. Y., Molina, L. T., and Molina, M. J.: The HCl+ClONO2 750

reaction rate on various water ice surfaces, Chem. Phys. Lett., 315, 7-11, 1999. 751

Liao, J., Huey, L. G., Liu, Z., Tanner, D. J., Cantrell, C. A., Orlando, J. J., Flocke, F. M., 752

Shepson, P. B., Weinheimer, A. J., Hall, S. R., Ullmann, K., Beine, H. J., Wang, Y., Ingall, E. 753

D., Stephens, C. R., Hornbrook, R. S., Apel, E. C., Riemer, D., Fried, A., Mauldin Iii, R. L., 754

Smith, J. N., Staebler, R. M., Neuman, J. A., and Nowak, J. B.: High levels of molecular 755

chlorine in the Arctic atmosphere, Nature Geosci., 7, 91-94, 2014. 756

McCormick, M. P., Thomason, L. W., and Trepte, C. R.: Atmospheric effects of the Mt 757

Pinatubo eruption, Nature, 373, 399-404, 1995. 758

Molina, L. T., Spencer, J. E., and Molina, M. J.: The rate constant for the reaction of O(3P) 759

atoms with ClONO2, Chem. Phys. Lett., 45, 158-162, 1977. 760

Molina, M. J., Molina, L. T., and Kolb, C. E.: Gas-phase and heterogeneous chemical 761

kinetics of the troposphere and stratosphere, Annu. Rev. Phys. Chem., 47, 327-367, 1996. 762

Molina, M. J., Molina, L. T., Zhang, R. Y., Meads, R. F., and Spencer, D. D.: The reaction of 763

ClONO2 with HCl on aluminum oxide, Geophys. Res. Lett., 24, 1619-1622, 1997. 764

Morgenstern, O., Braesicke, P., O'Connor, F. M., Bushell, A. C., Johnson, C. E., Osprey, S. 765

M., and Pyle, J. A.: Evaluation of the new UKCA climate-composition model – Part 1: The 766

stratosphere, Geosci. Model Dev., 2, 43-57, 2009. 767

O'Connor, F. M., Johnson, C. E., Morgenstern, O., Abraham, N. L., Braesicke, P., Dalvi, M., 768

Folberth, G. A., Sanderson, M. G., Telford, P. J., Voulgarakis, A., Young, P. J., Zeng, G., 769

Collins, W. J., and Pyle, J. A.: Evaluation of the new UKCA climate-composition model – 770

Part 2: The Troposphere, Geosci. Model Dev., 7, 41-91, 2014. 771

Osthoff, H. D., Roberts, J. M., Ravishankara, A. R., Williams, E. J., Lerner, B. M., 772

Sommariva, R., Bates, T. S., Coffman, D., Quinn, P. K., Dibb, J. E., Stark, H., Burkholder, J. 773

B., Talukdar, R. K., Meagher, J., Fehsenfeld, F. C., and Brown, S. S.: High levels of nitryl 774

chloride in the polluted subtropical marine boundary layer, Nature Geosci., 1, 324-328, 2008. 775

Phillips, G. J., Tang, M. J., Thieser, J., Brickwedde, B., Schuster, G., Bohn, B., Lelieveld, J., 776

and Crowley, J. N.: Significant concentrations of nitryl chloride observed in rural continental 777

Europe associated with the influence of sea salt chloride and anthropogenic emissions, 778

Geophys. Res. Lett., 39, L10811, 10.1029/2012gl051912, 2012. 779

Pope, F. D., Braesicke, P., Grainger, R. G., Kalberer, M., Watson, I. M., Davidson, P. J., and 780

Cox, R. A.: Stratospheric aerosol particles and solar-radiation management, Nature Clim. 781

Change, 2, 713-719, 2012. 782

Renard, J. J., and Bolker, H. I.: The chemistry of chlorine monoxide (dichlorine monoxide), 783

Chem. Rev., 76, 487-508, 1976. 784


31

Riedel, T. P., Bertram, T. H., Crisp, T. A., Williams, E. J., Lerner, B. M., Vlasenko, A., Li, 785

S.-M., Gilman, J., de Gouw, J., Bon, D. M., Wagner, N. L., Brown, S. S., and Thornton, J. A.: 786

Nitryl Chloride and Molecular Chlorine in the Coastal Marine Boundary Layer, Environ. Sci. 787

Technol., 46, 10463-10470, 2012. 788

Robinson, G. N., Worsnop, D. R., Jayne, J. T., Kolb, C. E., and Davidovits, P.: 789

Heterogeneous uptake of ClONO2 and N2O5 by sulfuric acid solutions, J. Geophys. Res.-790

Atmos., 102, 3583-3601, 1997. 791

Romanias, M. N., El Zein, A., and Bedjanian, Y.: Heterogeneous Interaction of H2O2 with 792

TiO2 Surface under Dark and UV Light Irradiation Conditions, J. Phys. Chem. A, 116, 8191-793

8200, 2012. 794

Rouviere, A., Sosedova, Y., and Ammann, M.: Uptake of Ozone to Deliquesced KI and 795

Mixed KI/NaCl Aerosol Particles, J. Phys. Chem. A, 114, 7085-7093, 2010. 796

Schack, C. J., and Lindahl, C. B.: On the synthesis of chlorine monoxide, Inorg. Nucel. 797

Chem. Lett., 3, 387-389, 1967. 798

Seisel, S., Borensen, C., Vogt, R., and Zellner, R.: Kinetics and Mechanism of the Uptake of 799

N2O5 on Mineral Dust at 298 K, Atmos. Chem. Phys., 5, 3423-3432, 2005. 800

Shang, J., Li, J., and Zhu, T.: Heterogeneous Reaction of SO2 on TiO2 Particles, Sci. China 801

Chem., 53, 2637-2643, 2010. 802

Shepherd, J.: Geoengineering the climate: science, governance and uncertainty, The Royal 803

Society, London, UK, 2009. 804

Shi, Q., Jayne, J. T., Kolb, C. E., Worsnop, D. R., and Davidovits, P.: Kinetic model for 805

reaction of ClONO2 with H2O and HCl and HOCl with HCl in sulfuric acid solutions, J. 806

Geophys. Res.-Atmos., 106, 24259-24274, 2001. 807

Simpson, W. R., Brown, S. S., Saiz-Lopez, A., Thornton, J. A., and Glasow, R. v.: 808

Tropospheric Halogen Chemistry: Sources, Cycling, and Impacts, Chem. Rev., 115, 4035-809

4062, 2015. 810

Solomon, S.: Stratospheric ozone depletion: A review of concepts and history, Rev. 811

Geophys., 37, 275-316, 1999. 812

SPARC: Assessment of Stratospheric Aerosol Properties (SPARCR Report No. 4), 2006. 813

Spicer, C. W., Chapman, E. G., Finlayson-Pitts, B. J., Plastridge, R. A., Hubbe, J. M., Fast, J. 814

D., and Berkowitz, C. M.: Unexpectedly high concentrations of molecular chlorine in coastal 815

air, Nature, 394, 353-356, 1998. 816

Stull, D. R.: Vapor Pressure of Pure Substances. Organic and Inorganic Compounds, Ind. 817

Eng. Chem., 39, 517-540, 1947. 818

Tang, M. J., Thieser, J., Schuster, G., and Crowley, J. N.: Kinetics and Mechanism of the 819

Heterogeneous Reaction of N2O5 with Mineral Dust Particles, Phys. Chem. Chem. Phys., 14, 820

8551-8561, 2012. 821

Tang, M. J., Camp, J. C. J., Rkiouak, L., McGregor, J., Watson, I. M., Cox, R. A., Kalberer, 822

M., Ward, A. D., and Pope, F. D.: Heterogeneous Interaction of SiO2 with N2O5: Aerosol 823

Flow Tube and Single Particle Optical Levitation-Raman Spectroscopy Studies, J. Phys. 824

Chem. A, 118, 8817-8827, 2014a. 825

Tang, M. J., Cox, R. A., and Kalberer, M.: Compilation and Evaluation of Gas Phase 826

Diffusion Coefficients of Reactive Trace Gases in the Atmosphere: Volume 1. Inorganic 827

Compounds, Atmos. Chem. Phys., 14, 9233-9247, 2014b. 828

Tang, M. J., Schuster, G., and Crowley, J. N.: Heterogeneous Reaction of N2O5 with Illite 829

and Arizona Test Dust Particles, Atmos. Chem. Phys., 14, 245-254, 2014c. 830

Tang, M. J., Telford, P. J., Pope, F. D., Rkiouak, L., Abraham, N. L., Archibald, A. T., 831

Braesicke, P., Pyle, J. A., McGregor, J., Watson, I. M., Cox, R. A., and Kalberer, M.: 832

Heterogeneous reaction of N2O5 with airborne TiO2 particles and its implication for 833

stratospheric particle injection, Atmos. Chem. Phys., 14, 6035-6048, 2014d. 834


32

Tang, M. J., Cziczo, D. J., and Grassian, V. H.: Interactions of Water with Mineral Dust 835

Aerosol: Water Adsorption, Hygroscopicity, Cloud Condensation and Ice Nucleation, Chem. 836

Rev., 116, 4205–4259, 2016. 837

Telford, P., Braesicke, P., Morgenstern, O., and Pyle, J.: Reassessment of causes of ozone 838

column variability following the eruption of Mount Pinatubo using a nudged CCM, Atmos. 839

Chem. Phys., 9, 4251-4260, 2009. 840

Telford, P. J., Braesicke, P., Morgenstern, O., and Pyle, J. A.: Technical Note: Description 841

and assessment of a nudged version of the new dynamics Unified Model, Atmos. Chem. 842

Phys., 8, 1701-1712, 2008. 843

Textor, C., Schulz, M., Guibert, S., Kinne, S., Balkanski, Y., Bauer, S., Berntsen, T., Berglen, 844

T., Boucher, O., Chin, M., Dentener, F., Diehl, T., Easter, R., Feichter, H., Fillmore, D., 845

Ghan, S., Ginoux, P., Gong, S., Grini, A., Hendricks, J., Horowitz, L., Huang, P., Isaksen, I., 846

Iversen, I., Kloster, S., Koch, D., Kirkevåg, A., Kristjansson, J. E., Krol, M., Lauer, A., 847

Lamarque, J. F., Liu, X., Montanaro, V., Myhre, G., Penner, J., Pitari, G., Reddy, S., Seland, 848

Ø., Stier, P., Takemura, T., and Tie, X.: Analysis and Quantification of the Diversities of 849

Aerosol Life Cycles within AeroCom, Atmos. Chem. Phys., 6, 1777-1813, 2006. 850

Thornton, J. A., Kercher, J. P., Riedel, T. P., Wagner, N. L., Cozic, J., Holloway, J., S., Dube, 851

W. P., Wolfe, G. M., Quinn, P. K., Middlebrook, A. M., Alexander, B., and Brown, S. S.: A 852

large atomic chlorine source inferred from mid-continental reactive nitrogen chemistry, 853

Nature, 464, 271-174, 2010. 854

Tilmes, S., Muller, R., and Salawitch, R.: The sensitivity of polar ozone depletion to 855

proposed geoengineering schemes, Science, 320, 1201-1204, 2008. 856

Tilmes, S., Mills, M. J., Niemeier, U., Schmidt, H., Robock, A., Kravitz, B., Lamarque, J. F., 857

Pitari, G., and English, J. M.: A new Geoengineering Model Intercomparison Project 858

(GeoMIP) experiment designed for climate and chemistry models, Geosci. Model Dev., 8, 859

43-49, 2015. 860

Wagner, C., Hanisch, F., Holmes, N., de Coninck, H., Schuster, G., and Crowley, J. N.: The 861

interaction of N2O5 with mineral dust: aerosol flow tube and Knudsen reactor studies, Atmos. 862

Chem. Phys., 8, 91-109, 2008. 863

Wang, T., Tham, Y. J., Xue, L., Li, Q., Zha, Q., Wang, Z., Poon, S. C. N., Dubé, W. P., 864

Blake, D. R., Louie, P. K. K., Luk, C. W. Y., Tsui, W., and Brown, S. S.: Observations of 865

nitryl chloride and modeling its source and effect on ozone in the planetary boundary layer of 866

southern China, J. Geophys. Res.-Atmos, 121, 2476-2489, 2016. 867

Weisenstein, D. K., Keith, D. W., and Dykema, J. A.: Solar geoengineering using solid 868

aerosol in the stratosphere, Atmos. Chem. Phys., 15, 11835-11859, 2015. 869

Wilson, J. C., Jonsson, H. H., Brock, C. A., Toohey, D. W., Avallone, L. M., Baumgardner, 870

D., Dye, J. E., Poole, L. R., Woods, D. C., Decoursey, R. J., Osborn, M., Pitts, M. C., Kelly, 871

K. K., Chan, K. R., Ferry, G. V., Loewenstein, M., Podolske, J. R., and Weaver, A.: In-situe 872

observations of aerosol and chlorine monoxide after the 1991 eruption of Mount-Pinatubo: 873

effect of reactions on sulfate aerosol, Science, 261, 1140-1143, 1993. 874

Wu, L.-Y., Tong, S.-R., Hou, S.-Q., and Ge, M.-F.: Influence of Temperature on the 875

Heterogeneous Reaction of Formic Acid on α-Al2O3, J. Phys. Chem. A, 116, 10390-10396, 876

2012. 877

Zhang, R. Y., Jayne, J. T., and Molina, M. J.: Heterogeneous interacyions of ClONO2 and 878

HCl with surlfuric acid tetrehydrate: implications for the stratosphere, J. Phys. Chem., 98, 879

867-874, 1994. 880

881

882


33

Tables& Figures 883

Table 1. Loss rates and uptake coefficients of ClONO2 onto the inner wall of the Pyrex tube at 884

different relative humidities (RH). 885

RH (%) kw (×10-2 s-1) γ (×10-6)

0 3.6±0.2 5.1±0.3

2.9±0.4 3.9±0.6

6 4.1±0.1 6.2±0.1

3.7±0.7 5.4±1.0

12 4.1±0.3 6.2±0.5

17 6.9±0.3 13±0.6

6.4±0.2 11±0.4

24 8.1±0.8 16±2.0

8.2±0.3 17±0.7

886


34

Table 2. Uptake coefficients of ClONO2 onto SiO2 and TiO2 aerosol particles at different 887

relative humidities (RH). ka: loss rate of ClONO2 onto aerosol particle surface; Sa: aerosol 888

surface area concentration; γ(ClONO2): uptake coefficients of ClONO2. 889

Particle RH (%) ka

(×10-3 s)

Sa

(×10-3 cm2 cm-3)

γ(ClONO2)

(×10-4)

SiO2 7±1 4.1±2.5 2.80±0.02 2.3±1.4

7±1 3.4±3.2 2.78±0.05 1.9±1.8

17±2 <5.1 a 1.08±0.08 <7.5 a

17±2 <5.4 a 1.28±0.07 <6.7 a

17±2 <7.3 a 1.78±0.09 <6.5 a

17±2 6.5±4.2 2.08±0.06 4.9±3.2

35±4 6.3±3.1 2.34±0.08 4.2±2.1

35±4 13.1±4.7 2.91±0.09 7.1±2.6

35±4 9.0±7.3 2.86±0.10 4.8±3.9

59±3 11.6±3.5 2.88±0.06 6.4±1.9

TiO2 7±1 7.0±1.4 1.09±0.12 10.1±2.0

7±1 6.2±2.3 0.73±0.05 13.7±5.0

33±3 17.9±5.6 2.23±0.03 12.7±3.9

33±3 14.5±1.4 1.93±0.03 11.9±1.1

a: estimated upper limits. 890


35

891

Figure 1. Schematic diagram of the aerosol flow tube used in this study. SMPS: Scanning 892

Mobility Particle Sizer; CLD: Chemiluminescence detector, used to measure the ClONO2 893

concentration (measured as the change in NO concentration). All the flows (except the flow 894

applied to the atomizer) were controlled by mass flow controllers. Flow details are provided in 895

text. 896


36

897

Figure 2. Response of measured NO and NOx mixing ratios to the introduction of ClONO2 898

into the flow tube (left panel). The corresponding calculated ClONO2 mixing ratio and nitrogen 899

balance are also shown (right panel). 900


37

901

Figure 3. Decays of ClONO2 in the flow tube due to its loss onto the Pyrex glass (circles: 0% 902

RH; squares: 24% RH). Measured ClONO2 mixing ratios were normalized to that at 8.5 s (when 903

the injector was at 30 cm). Typical ClONO2 mixing ratios in the flow tube are a few hundred 904

ppbv (see Figure 2). 905


38

906

Figure 4. Decays of ClONO2 in the aerosol flow tube without (open circles) and with (solid 907

squares) aerosol particles in the aerosol flow tube under different experimental conditions. (a) 908

TiO2 with a surface area concentration of 2.3×10-3 cm2 cm-3 at 33% RH; (b) SiO2 with a surface 909

area concentration of 2.9×10-3 cm2 cm-3 at 39% RH. 910


39

911

Figure 5. Dependence of γ(ClONO2) on RH for Pyrex glass. 912


40

913

914 Figure 6. Simulated annual mean, zonal mean N2O5 percentage differences between TiO2 915

injection (S3) and the Mt Pinatubo eruption (S2). Black contour lines show N2O5 mixing ratios 916

from the Mt Pinatubo simulation (S2) in ppb. 917


41

918

919

Figure 7. Simulated annual mean, zonal mean ClOx percentage differences between TiO2 920

injection (S3) and the Mt Pinatubo eruption (S2). Black contour lines show ClOx mixing ratios 921

from the Mt Pinatubo simulation (S2) in ppb. 922


42

923

Figure 8. Simulated annual mean, zonal mean O3 percentage differences between TiO2 924

injection (S3) and the Mt Pinatubo eruption (S2). Black contour lines show ClOx mixing ratios 925

from the Mt Pinatubo simulation (S2) in ppmv. 926


heterogeneous reaction of clono with tio and sio aerosol...

Documents